Progress on China nuclear data processing code system
Ping Liu1,a, Xiaofei Wu1, Zhigang Ge1, Songyang Li2, Haicheng Wu1, Lili Wen1, Wenming Wang1, and Huanyu Zhang1 1 China Nuclear Data Center, China Institute of Atomic Energy, Beijing 102413, China
2 Tsinghua University, Beijing 10086, China
Abstract. China is developing the nuclear data processing code Ruler, which can be used for producing multi-group cross sections and related quantities from evaluated nuclear data in the ENDF format [1]. The Ruler includes modules for reconstructing cross sections in all energy range, generating Doppler-broadened cross sections for given temperature, producing effective self-shielded cross sections in unresolved energy range, calculating scattering cross sections in thermal energy range, generating group cross sections and matrices, preparing WIMS-D format data files for the reactor physics code WIMS-D [2]. Programming language of the Ruler is Fortran-90. The Ruler is tested for 32-bit computers with Windows-XP and Linux operating systems. The verification of Ruler has been performed by comparison with calculation results obtained by the NJOY99 [3] processing code. The validation of Ruler has been performed by using WIMSD5B code.
1. Introduction
The Ruler is a computer code package for generating multi-group cross sections and related quantities from evaluated nuclear data in the ENDF format.
The objective to create the computer programs for the production of multi-group constants for the calculation of reactor physics using evaluated data was conceived 6 years ago. By that time, the project of the self-reliance development of key design codes for nuclear power plants was created, and the self-reliance development of nuclear data processing code is an essential part one. The Ruler has been developed since then.
The Ruler has been developed through the formula derivation, requirements analysis, system design, detailed design, coding, unit testing, sub-system testing and system testing. The well-proven models and methods for generating multi-group constants are used, including resonance reconstruction, Doppler broadening, thermal scattering processing, group cross section and scattering matrices computing, as well as resonance self-shielding treatment.
The Ruler consists of a set of functional modules, each performing a defined processing task. They are main code, data management modules, physics modules, interface modules. They are linked to one another by input and output files. Each of functional modules is supported by a number of subsidiary modules.
The Ruler is China’s first multi-group constants generating code system for reactor physics and shielding design, and the development of Ruler will be continued.
2. Content of the Ruler
The Ruler is designed with a modular structure, which contains four functional modules, which are supported by a number of subsidiary modules. They are the main code, ae-mail: [email protected]
the data management modules, the physics modules, and the interface modules.
2.1. Architecture of the ruler
For efficient programming, we used the independent ENDF I/O system in the development of the Ruler code, that the data input and output module is separated from the computation modules in the Ruler.
The Ruler has four functional modules, which consists of 11 subsidiary modules, they are the main driver module, data input and output module, public mathematics module, system common subroutine module, data mode convert-ing module, resonance reconstruction module, Doppler broadening module, the unresolved resonance self-shielding module, thermal data generating module, neu-tron multi-group constants producing module, WIMS-D format data library preparing module, respectively.
Figure 1 shows the architecture of Ruler. The data management modules, the physics modules and the interface modules can be used with relatively independence. Those modules can access the PENDF files obtained by other processing code.
2.2. The endf input-output system
The Ruler has independent ENDF I/O system, which consists of two parts. They are the data type and user interface. The data type includes all ENDF files, such as cross sections, angular distribution etc. The user interface includes reading ENDF files and writing ENDF files etc.
Table 1 gives the user interface and function of ENDF IO.
The basic concept of ENDF I/O system is packaging the decentralized I/O operation into a library that programming-maker can use API (Application Program-ming Interface)-Centric programProgram-ming pattern to replace the File-centric programming pattern of NJOY.
Figure 1. Structure of the Ruler.
Figure 2. Different programming patterns between Ruler and NJOY.
Table 1. User interface and function of ENDF IO.
Interface Function
Subroutine Read data from read endf file(fin,endf,mat) (P)ENDF files
Subroutine Read data from read gendf file(fin,endf,mat) GENDF files
Subroutine
Write onto a write endf file(fout,endf,
(P)ENDF files overwrite)
Subroutine
Write onto a write gendf file(fout,endf,
GENDF files overwrite)
Subroutine del endf(endf) Release memory space Subroutine del gendf(gendf) Release memory space
Interface insert mt Insert file onto specified position Interface find mt Find specified MT file
Figure 2 shows the comparison of programming patterns between Ruler and NJOY.
The use of independent ENDF I/O system in the Ruler results that the Ruler is much more advanced in the aspects of the running time, the maintainability and the extendibility, especially is easy for modification while the ENDF format is changed.
Figure 3. Flow diagram of the Ruler.
2.3. Flow diagram of the ruler
The Ruler can be used for generating the WIMS-D format data files from ENDF format files. Figure3shows the flow diagram of the Ruler.
The run of the Ruler begins with the reading of ENDF format files. Then, a successive data processing for each module is carried out.
The following brief summary illustrates the general flow of data processing in the Ruler.
MODE converts ENDF library files for binary mode for calculation efficiency.
RXSP reads an ENDF file and produces a common energy grid for all reactions. Resonance cross sections are reconstructed in all energy range. The resulting pointwise cross sections are written onto a “point-ENDF” (PENDF) file for future use.
DOPLER reads a PENDF file and Doppler-broadens the data using the accurate point-kernel method to generate Doppler-broadened cross sections for given temperature. The results are written out on a new PENDF file for future use.
NURD produces effective self-shielded pointwise cross sections, versus temperatures and background cross sections in the unresolved range. This is done for each temperature produces by DOPLER. The results are added to the PENDF file.
TSC produces scattering cross sections in the thermal range. The results are added to the PENDF file.
GAC processes the pointwise cross sections into multigroup form. The weighting function for group averaging can be taken to be the Bondarenko form, or it can be computed from the slowing-down equation for a heavy absorb in a light moderator. The results are saved on a GENDF file.
GWIMS prepares libraries for the thermal reactor assembly code WIMS-D.
3. Verification and validation of Ruler
The verification and validation of Ruler has been performed by comparison with the calculation results obtained by NJOY99 code.Figure 4. Reconstruction fission cross sections for U-235.
Figure 5. Reconstruction fission cross sections for Pu-241.
Figure 6. Doppler broadening fission cross sections for U-235.
3.1. Comparison of calculated cross sections
Figures 4 to 11 show the comparison of cross sections for some materials between Ruler and NJOY. It can be seen that the reconstruction cross sections of Ruler show agreement with NJOY99 within 1, Doppler broadening cross sections of Ruler show agreement with NJOY99 within 3, group averaging cross sections are similar for
Figure 7. Doppler broadening fission cross sections for U-235.
Figure 8. 69 groups averaged fission cross sections for U-235.
Figure 9. 69 groups averaged fission cross sections for Pu-241.
Ruler and NJOY, and the WIMS-D format cross sections of Ruler show agreement with NJOY99 within 1.
3.2. Comparison of running time
Table 2 shows the comparison of running time for generating a WIMS-D format data file between Ruler and NJOY99. It can be seen that Ruler has less computing time compared with NJOY99. This is because of the independent ENDF I/O system in the Ruler that read-once,
10-5 10-3 10-1 101 103 105 107 0.99996 0.99998 1 1.00002 1.00004 1.00006 1.00008 1.0001 Cr os s Sectio n Ra ti o (Ru ler/ N J O Y) Energy (eV) ABSORPTION TRANSPORT FISS. YIELD FISSION SCATTERING T=293K
Figure 10. WIMS-D 69 groups cross sections for U-235.
10-5 10-4 10-3 10-2 10-1 100 101 102 103 104 105 106 107 0.99997 0.99998 0.99999 1 1.00001 1.00002 1.00003 1.00004 1.00005 1.00006 1.00007 1.00008 1.00009 1.0001 Cr o s s Sec tio n Ra tio (Ruler/NJ OY) Energy (eV) ABSORPTION TRANSPORT FISS. YIELD FISSION SCATTERING T=293K
Figure 11. WIMS-D 69 groups cross sections for Pu-239.
Table 2. Comparison of running time. Nuclide Running time (s) Speed-up ratio
NJOY Ruler NJOY/Ruler 1-H-H2O 152.7 104.2 1.47 1-H-ZrH 410.2 365.9 1.12 1-D-D2O 157.6 101.6 1.55 6-C-nat 84.4 61.7 1.37 8-O-16 25.9 14.9 1.74 13-Al-27 7.0 4.3 1.63 26-Fe-56 23 19.9 1.16 36-Kr-83 5.7 3.1 1.84 42-Mo-95 7 5.7 1.23 92-U-235 247.3 187.0 1.32 92-U-238 448.6 225.3 1.99 94-Pu-239 136.7 67 2.04
use-many in the Ruler, but read-many, use-many in the NJOY.
3.3. Benchmark tests for Ruler
Benchmark tests for Ruler have been performed using the WIMSD5B code. The ENDF files are from ENDF/ B-VII.1 [4]. All benchmarks are from WLUP/IAEA [5]. The WIMS-D format data files were generated with Ruler according to the flow diagram of Fig.2.
Figures 12 to 13 show the comparison of bench-marking results between Ruler, NJOY and experiments. The results of NJOY are from WLUP/IAEA [6]. It can
0.30 0.35 0.40 0.45 0.50 0.55 0.60 0.65 0.70 0.75 0.80 0.980 0.984 0.988 0.992 0.996 1.000 1.004 1.008 1.012 Experiment Ruler NJOY keff Slowing-down desity
Figure 12. H2O moderated UO2 lattices ke f fvs. q. Enrichment (>2 and <4 wt% u-235). 0.4 0.5 0.6 0.7 0.8 0.9 0.985 0.990 0.995 1.000 1.005 1.010 1.015 1.020 1.025 Experiment Ruler NJOY kef f Slowing-down desity
Figure 13. H2O moderated MOX lattices ke f fvs. q. Enrichment (>2 wt% fissile metal).
be seen that the calculated results of Ruler show good agreement with those of NJOY99 for H2O moderated uranium oxide lattices. All values of ke f f generated with Ruler are slightly lower than the corresponding values obtained with NJOY99 H2O moderated MOX lattices, with similar tread for Ruler and NJOY.
4. Conclusions and intentions
The primary objective to create nuclear data processing code has been accomplished. WIMS-D format cross sections and other related quantities were generated using Ruler. The verification of Ruler has been performed by comparison with calculation results obtained by NJOY99.396 processing code. The benchmark tests for the library prepared by Ruler have been performed with WIMSD5B code.
The future development of Ruler will be continued, which includes photon processing module and more interface modules. The modification of ENDF IO module for new data format will be considered.
References
[1] M. Herman and A. Trkov, Eds., ENDF-6 Formats Manual, Data Formats and Procedures for the evalu-ated nuclear data File ENDF/B-VI and ENDF/B-VII, Brookhaven National Laboratory report BNL-90365-2009 (ENDF-102) (June BNL-90365-2009). For the latest version of this report, see http//www.nndc.bnl.gov and take the ENDF tab to ENDF-6 formats Manual
[2] WIMS-D5, OECD/NEA Data Bank
Documentation, Package ID No. 1507/02 (1998), http//www.nea.fr/html/dbprog/
[3] R. E. MacFarlane and D. W. Muir. NJOY99.0, Code System for Producing Pointwise and Multigroup Neutron and Photon Cross Sections from ENDF/B Data. Los Alamos National Laboratory Report, LA-12740-M, 2000
[4] M. B. Chadwick et al., “ENDF/B-VII.1 Nuclear Data for Science and Technology : Cross sections, Covariances, Fission Product Yields and Decay Data,” Nuclear Data Sheets 112, 2887–2996 (2011)
[5] http://www-nds.iaea.org/wimsd/benchmarks.
htm
[6] http://www-nds.iaea.org/wimsd/results/
